Vitaly Buka

Results 10 issues of Vitaly Buka

It's used in btVector3::setMax

Default constructor of btVector3 does not initialize the object. This fixes Memory Sanitizer reports.

DeleteHandler must outlive the the first DeleteHandlerPointer. ApplicationUsage from ApplicationUsage::instance() was created by ApplicationUsage.cpp module constructor and the unstecified order related to module constructor of Referenced.cpp which creates DeleteHandlerPointer. Then...

Here us the documentation: https://google.github.io/googletest/advanced.html#global-set-up-and-tear-down It says: RUN_ALL_TESTS() always calls TearDown() with each environment object, regardless of whether or not the tests were run. However the code is different: https://github.com/google/googletest/blob/5126f7166109666a9c0534021fb1a3038659494c/googletest/src/gtest.cc#L5847-L5903

bug

https://lab.llvm.org/staging/#/builders/sanitizer-x86_64-linux-bootstrap-msan

For #250. This bots can't catch-up with `clean_obj` requests, but they do clean build anyway. Maybe we should do the same for all sanitizer builders, but seems they are OK...

Follow up to #2922 and #2917. Before #2922 the only possible "else" with "opt" was space separated flags. Now we get there for already `is_absolute` path.

According to https://github.com/llvm/llvm-project/pull/115361#issuecomment-2477500295 unrelated steps of current bootstrap bots unnecessarily spam libc++ patches.

Looks like a regular tag mismatch even on commits which were green before. No known changes in bot setup or zorg script. ``` Testing: 0.. 10.. 20.. 30.. 40.. 50.....

Detected with ``` cmake .. -GNinja \ -DCMAKE_C_FLAGS="-fsanitize=pointer-overflow -fno-sanitize-trap=pointer-overflow -fno-sanitize-recover=pointer-overflow" \ -DCMAKE_CXX_FLAGS="-fsanitize=pointer-overflow -fno-sanitize-trap=pointer-overflow -fno-sanitize-recover=pointer-overflow" \ -DCMAKE_EXE_LINKER_FLAGS="-fsanitize=pointer-overflow" -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++ ninja ninja test ```